Skiba points in free end-time problems
Introduction
Sethi (1977), Sethi (1979), Skiba (1978), Dechert and Nishimura (1983) established the existence of indifference points, or Skiba points,1 in infinite time optimal control models with multiple steady states. At a Skiba point the decision maker is indifferent between choosing trajectories that approach different steady states.
The existence of Skiba points implies that the optimal solution is history dependent, i.e. which long run steady state will be reached depends on the initial value of the state variable. For policy makers such knowledge is especially important when one steady state can be designated as a “bad” equilibrium, while the other one is a “good” equilibrium. Then, by implementing the right policy measure, a policy maker can shift the Skiba point in such a way that the “good” equilibrium will be reached. Papers on this topic include (Matsuyama, 1991, Hartl and Kort, 1996) on an environmental economics application, and Caulkins et al. (2013) on rational addiction.
This paper extends the notion of Skiba points to the class of optimal control problems in which the decision maker can stop operations at any time and collect a salvage value. In such a framework essentially three qualitatively different solution candidates can be distinguished. Besides operating forever, the decision maker can decide to stop operations immediately and collect the salvage value, or to operate for a finite time and then stop. We show that Skiba points exist where the decision maker is indifferent (1) between “operating forever” and “operating for a finite time and then stopping”, and (2) between “operating forever” and “stopping immediately”.
We focus on Skiba points where the decision maker is indifferent between the policies “operate forever” or “operate for a finite time and collect the salvage value at the endogenous horizon date” since then both options involve dynamics. One might expect that in a one-dimensional model, if one policy involves departing the Skiba point by increasing the state variable, then for the other policy the state must start to decrease. We prove that is true – almost always. There is a hairline case in which both policies can move away from the Skiba point in the same direction, but in that case they apply the same control and all points traversed are Skiba points; the two alternate solutions move from one Skiba point to another Skiba point until reaching a point where one is indifferent between continuing to operate forever and ceasing operation immediately.
Recent contributions in the area of multiple optimal solutions include (Kato et al., 2007, Kiseleva and Wagener, 2010, Ernst and Semmler, 2010, Zeiler et al., 2010, Akao et al., 2011).
We illustrate our results by studying the problem of a company whose management has the option to sell the firm (be acquired). This problem could be relevant with regard to high growth, technology-oriented companies. Some such companies remain independent, and some grow to become very profitable (e.g. Google whose founders are now billionaires, Apple, Facebook, and Adobe Systems). Yet, for every Google, there are many other startups that make their founders rich in a different way, by allowing the founders to grow the technology for a time, but then be acquired voluntarily. Examples are Hotmail, Android, YouTube and PayPal (acquired by Microsoft, Google and eBay).
While there are many capital accumulation models (e.g., Jorgenson, 1963, Gould, 1968, Barucci, 1998, Feichtinger et al., 2006), to our knowledge none focus on the possibility that it could be optimal that a startup may want to be acquired by another (usually larger) firm.
The remainder of the paper is organized as follows. Section 2 shows for a general framework that whenever Skiba points exist where the decision maker is indifferent between a policy of operating forever or ceasing operation at a finite point of time, then if for the policy of staying alive forever the state variable starts by growing, exactly the opposite happens when the decision maker chooses to stop operations in finite time, and vice versa. Section 3 illustrates this theory by analyzing the capital accumulation model of the firm where the manager has the option to sell it at any time. Section 4 concludes.
Section snippets
Skiba points in a free end-time problem
This section characterizes the optimal paths in one-state free end-time problems, specifically when a Skiba point exists where the decision maker is indifferent between “operating forever” and “operating during a finite time interval” after which the decision maker collects the salvage value. At such a Skiba point it holds that if the state variable starts growing under one policy, it starts by shrinking for the other policy (see also Fig. 3, Fig. 5, Fig. 7 corresponding to our illustrative
The model of the firm
We start from a concave revenue function version of the capital accumulation model by Barucci (1998) and Hartl and Kort (2000), Hartl and Kort (2004). We extend this model by introducing the option that the start-up firm can be taken over by another, established company.
The firm starts with some goodwill which we model as a fixed constant, m, as it is not the focus of the dynamics under investigation here. In addition the firm has a capital stock, K, that it exploits to produce revenue and in
Conclusion
The starting point of this paper is the literature on Skiba points, including the early contributions by Sethi (1977), Sethi (1979), Skiba (1978), Dechert and Nishimura (1983). This literature considers infinite time optimal control models with multiple steady states. A Skiba point is defined such that right at this point the decision maker is indifferent between choosing trajectories that approach different steady states. This paper generalizes the Skiba point literature to models with a free
Acknowledgements
The authors like to thank the editor, three anonymous reviewers and Vladimir Veliov for their helpful comments. This research was supported by the Austrian Science Fund (FWF) under Grant P21410-G16 and P23084-N13.
References (22)
- et al.
Monotonicity and continuity of the critical capital stock in the Dechert-Nishimura model
J. Math. Econ.
(2011) - et al.
A complete characterization of optimal growth paths in an aggregated model with a non-concave production function
J. Econ. Theory
(1983) - et al.
Global dynamics in a model with search and matching in labor and capital markets
J. Econ. Dyn. Control
(2010) - et al.
Anticipation effects of technological progress on capital accumulation: A vintage capital approach
J. Econ. Theory
(2006) A simple proof of the monotonicity of the state trajectories in autonomous control problems
J. Econ. Theory
(1987)- et al.
Bifurcations of optimal vector fields in the shallow lake model
J. Econ. Dyn. Control
(2010) Optimal investments with increasing returns to scale
Int. Econ. Rev.
(1998)- et al.
Differential Equations, Stability and Chaos in Dynamic Economics
(1989) - et al.
Multiple equilibria and indifference-threshold points in a rational addiction model
Central Eur. J. Oper. Res.
(2013) Adjustment costs in the theory of investment of the firm
Rev. Econ. Stud.
(1968)
Optimal Control of Nonlinear Processes: With Applications in Drugs, Corruption and Terror
Cited by (16)
A causal structure-based multiple-criteria decision framework for evaluating the water-related ecosystem service tradeoffs in a desert oasis region
2022, Journal of Hydrology: Regional StudiesCitation Excerpt :In general, the decision makers would be indifferent intermediate level of recommended values between two alternatives, which can shift the Skiba point in such a way that the Optimal recommended value will be reached. This also verifies the existence of Skiba point in ecosystem service modeling (Friedlob and Ramsay, 1986; Scott and Antonsson, 2000; Wagener, 2003; Caulkins et al., 2015). With agricultural water use accounting for 97.7% of the total water supply in this arid oasis (Xue et al., 2017a), improving the efficiency of water use via water-saving measures is crucial in alleviating conflicts over water in regional sustainable development.
Why (some) abnormal problems are “normal”
2021, Systems and Control LettersThe optimal lockdown intensity for COVID-19
2021, Journal of Mathematical EconomicsCitation Excerpt :However, both of these papers consider infinite time horizon models, whereas in our framework the time horizon is finite. In that sense the model of Caulkins et al. (2015) is more related, but there just Skiba points in the usual sense, i.e. separating “only” two different solutions with equal objective value, occur. We proceed by introducing the model.
On the Matthew effect in research careers
2021, Journal of Economic Dynamics and ControlCitation Excerpt :Bifurcations associated with the occurrence of Skiba points are studied in Wagener (2003), Kiseleva and Wagener (2010). Skiba points in free end-time problems are discussed in Caulkins et al. (2015), Seidl et al. (2016). By the introduction of the concept of the Stalling Equilibrium, we extend the literature on history-dependent solutions.
Optimal scientific production over the life cycle
2019, Journal of Economic Dynamics and ControlCitation Excerpt :Seidl et al. (2016) extend their approach by considering the option to leave academia. To do so, they compare, similar to the approach analyzed in Caulkins et al. (2015), solutions with a free end time (i.e. solutions, where it is optimal to leave academia) to infinite time horizon solutions (i.e. solutions, where it is optimal to stay in academia). They find that the optimal long-run career strategy can be history-dependent, i.e. it depends on the initial skills in research and teaching.
Delaying product introduction: A dynamic analysis with endogenous time horizon
2019, Journal of Economic Dynamics and Control